Urban operation index monitoring method based on multi-source heterogeneity

Abstract

The invention provides an urban operation index monitoring method based on multi-source heterogeneity, and the method comprises the steps: determining an urban operation monitoring target, determining a field needing to be monitored, carrying out the design of a target monitoring index, analyzing a data source needed by an attention index, converging urban operation multi-source data, and carrying out data analysis and standardization processing. The method comprises the following steps: setting a data sample, establishing an analysis model, carrying out model training based on historical data, and constructing an urban operation index monitoring model with adaptive capacity through rule setting and model parameter adjustment and optimization. According to the method, data resources in the fields of industrial economy, urban governance, green ecology, happiness and livability and the like produced in the urban operation process are fused, and data decision support is provided for the urban operation comprehensive situation and sensing of urban operation risks and development trends through data aggregation, data mapping, data governance, model analysis and other operations.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban governance, and in particular relates to a method for monitoring urban operation indicators based on multi-source heterogeneity. Background Art

[0002] With the rapid development of urbanization, the complexity of urban operation management is increasing. A large amount of data is generated in the fields of ecological environment, public safety, economic operation, etc. involved in urban operation, which has the characteristics of wide sources, diverse types, and different structural types. In recent years, the rapid development of technologies such as big data and artificial intelligence has provided new technical means for processing multi-source heterogeneous data, which can quickly realize the fusion, analysis and prediction of urban operation status based on multi-source heterogeneous data.

[0003] The existing urban operation indicator monitoring methods mainly focus on a single data source or a specific type of data and present it after simple calculation, which cannot fully reflect the real situation of urban operation. It is urgent to build an urban operation indicator monitoring method that can process multi-source heterogeneous data to realize urban situation monitoring, urban event perception and urban situation analysis, and perceive urban operation risks and development trends. Summary of the invention

[0004] Purpose of the invention: The technical problem to be solved by the present invention is to provide a city operation indicator monitoring method based on multi-source heterogeneity in view of the shortcomings of the existing technology, so as to realize the aggregation, integration and governance of heterogeneous data resources between different departments and industries of urban operation, and systematically quantify the city governance status.

[0005] Technical solution: A method for monitoring urban operation indicators based on multi-source heterogeneity, comprising the following steps:

[0006] Step 1: Determine the city operation monitoring objectives and design monitoring indicators;

[0007] Step 2: Analyze monitoring indicators to determine data sources and conduct multi-source data aggregation and management;

[0008] Step 3: Build a city operation indicator monitoring model with adaptive capabilities;

[0009] Step 4: Carry out monitoring and application of urban operation indicators in a hierarchical and classified manner.

[0010] Step 1 includes:

[0011] Step 1-1, conduct a panoramic scan of urban operations around the scenarios of economic governance, urban governance, public safety, etc., and construct indicators that can dynamically monitor the status of urban operations, such as designing economic indicators such as GDP growth rate, per capita disposable income, growth rate of the number of registered enterprises, tax revenue, etc., to monitor the overall status and trends of urban economic development.

[0012] Step 1-2: For different application scenarios, multidimensional analysis is used to design different types of indicators, including quantitative indicators, qualitative indicators, trend indicators and comparative indicators, to form a basic indicator library for daily monitoring of urban operations.

[0013] Step 1-3, determine a unique indicator for each record in the indicator library. The indicator library consists of three parts: indicator code, indicator name and indicator attribute. The indicator code is the unique identifier of the indicator; the indicator name is the Chinese name or English abbreviation of the indicator, which can uniquely identify a certain indicator and store data indicators; the indicator attribute is a detailed definition of the connotation and extension of the indicator, including indicator explanation, calculation method, logical relationship, accuracy and other contents.

[0014] Steps 1-4, based on the planning and classification of indicators and data, formulate a complete standardized management process for the indicator system, which can uniformly manage the changes of indicators to be released, review and release, indicator data access methods, online training, etc. At the same time, technical means such as data interfaces and APIs are used to realize the automatic acquisition and change of data;

[0015] Steps 1-5 use a four-level system to design indicators, covering first-level fields, second-level topics, third-level indicators, and fourth-level data items. For each indicator, the overall process and technical path for data collection, processing, analysis and display are included, including at least the indicator name, indicator unit, update frequency, responsible department, alarm threshold, target threshold, etc.

[0016] Step 2 includes:

[0017] Step 2-1, use multiple data collection methods such as API interface, direct database connection, file transfer, etc. to ensure the diversity and flexibility of data sources. During the data collection process, implement data integrity, consistency, and accuracy checks, such as field non-null checks, data type matching, data range verification, etc. According to the indicator requirements and data source characteristics, set a reasonable data collection frequency, such as real-time collection, daily collection, weekly collection, etc.;

[0018] Step 2-2, use statistical software, data mining tools, etc. to conduct correlation analysis and cluster analysis on indicators to identify repeatable and related indicators. Based on the analysis results, formulate clear screening and merging strategies, such as retaining the most representative indicators, calculating comprehensive indicators, etc.;

[0019] Step 2-3: Conduct research and collect opinions, conduct statistical analysis and text mining on the collected research data, optimize the selected indicators based on the opinions and suggestions, and form a set of practical and operational indicator system;

[0020] Step 2-4: Establish a data traceability mechanism to obtain the entire chain of information from the source to the mapping indicator, including the source of the data, the processing process, the analysis results, etc., and provide tools such as data query, report generation, and data visualization to facilitate managers to check and verify the accuracy and reliability of the data;

[0021] Step 2-5: Establish an update monitoring mechanism, formulate a data source update strategy that combines regular updates and on-demand updates, use timed tasks, triggers and other technologies to automate the update of indicator data, track data source updates in real time, and promptly discover and resolve problems that arise during the update process;

[0022] Steps 2-6: Use machine learning, natural language processing and other technical means to automatically label and classify multi-source data, automatically adapt labels according to data sources, build a list of associated mappings between indicators and data resources, establish a unified indicator visualization configuration function, and centrally manage and display the aggregated multi-source data;

[0023] Steps 2-7: Use ETL (Extract-Transform-Load), data warehouse and other technologies to manage the task of aggregating data, form tasks and automatically perform data aggregation, and implement data quality monitoring, data flow monitoring, system performance monitoring, etc. during the extraction, cleaning, transformation and loading of heterogeneous multi-source data. Simultaneously establish an exception handling mechanism to promptly respond to and handle abnormal situations that occur during data collection, processing and transmission;

[0024] Steps 2-8 support access control mechanisms based on roles and permissions, and use encryption algorithms such as AES (Advanced Encryption Standard) and RSA (Rivest-Shamir-Adleman, asymmetric encryption algorithm) to set security rules for the aggregated data, automatically identify sensitive data, encrypt and desensitize data, classify and grade data, perform minimum access control, data authorization and approval, and perform data security watermarking to establish a data security and trust space.

[0025] Step 3 includes:

[0026] Step 3-1: Clean the collected data, including removing duplicate data, processing missing values, correcting erroneous data, etc. For unstructured data (such as text, images, etc.), use analysis tools and technologies including OCR (Optical Character Recognition) and NLP (Natural Language Processing) to convert them into structured or semi-structured data;

[0027] Step 3-2: For the cleaned and parsed data, use dimensionality reduction techniques such as PCA (Principal Component Analysis) and t-SNE (t-Distributed Stochastic Neighbor Embedding) to reduce the dimension of the data, reduce the computational complexity and remove redundant information. At the same time, use the Z-score normalization method to convert data of different dimensions to the same dimension;

[0028] Step 3-3, based on business requirements and data characteristics, define the prototype of the monitoring indicator, including the indicator name, description, calculation formula, etc. Determine the data objects related to the indicator, including data source, data field, data type, etc.;

[0029] Step 3-4: Based on the indicator library prototype and data objects, select algorithm frameworks such as linear regression, decision tree, neural network, etc., preliminarily build the original model of the monitoring indicator, and set the initial parameters of the model;

[0030] Step 3-5: Set reasonable weights, thresholds, and evaluation criteria for the model based on business needs and data characteristics, and use historical data or simulated data to train the model to adjust the model parameters to achieve optimal performance;

[0031] Steps 3-6: Use machine learning, natural language analysis and other tools to build a knowledge graph to represent the associations and relationships between data, providing technical support for model optimization;

[0032] Steps 3-7: Use machine learning algorithms such as gradient descent, stochastic gradient descent, and Adam optimizer to tune the model parameters. During the iterative training process, the weights and bias items of the model are gradually optimized to improve the prediction accuracy of the model.

[0033] Step 3-8, using relationship extraction techniques such as rule-based methods and statistical methods to automatically extract relationship information related to specific indicators from multi-source heterogeneous data, and mapping the extracted relationship information with the indicator model to establish a relationship mapping table or relationship map;

[0034] Step 3-9: According to the relationship mapping table or relationship graph, the process of extracting related data is automatically orchestrated and managed, including process design, task scheduling, error handling, etc., to automatically extract related data related to specific indicators;

[0035] Step 3-10: Publish the verified and optimized indicator model online, and use visualization technologies such as data dashboards and three-dimensional simulation to output the model calculation results to form a panoramic view of urban operations.

[0036] Step 4 includes:

[0037] Follow steps 1 to 4 to design operation monitoring indicators, aggregate data and establish models for key areas of urban operation, such as industrial economy, urban governance, green ecology, and happy and livable environment, in a hierarchical, classified and field-based manner. The operation steps are similar.

[0038] The principle of the present invention is to carry out the design of operation monitoring indicators, data aggregation and model establishment for the city operation in a hierarchical, classified and field-based manner, formulate a complete indicator system standardization management process by building a basic indicator library and a source data collection and management library for daily monitoring of urban operation, adopt machine learning, natural language processing and other technical means to reduce the dimension and standardize multi-source heterogeneous data, select linear regression, decision tree, neural network and other algorithm frameworks, build the original model of monitoring indicators, establish the relationship mapping or relationship map between monitoring indicators and multi-source data, realize urban situation monitoring, urban event perception and urban situation analysis, and perceive urban operation risks and development trends.

[0039] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: the present invention adopts machine learning, natural language processing and other technical means to reduce the dimension and standardize the digitized data and unstructured data from different sources produced in urban operation, and automatically adapt the labels of data sources and indicator items; by constructing an operation monitoring indicator model, a relationship mapping or relationship map between monitoring indicators and multi-source data is established, and the model-related data corresponding to different indicators can be automatically updated and extracted, thereby solving the problem of inaccurate and incomplete urban operation monitoring under multi-source heterogeneous data. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more clear.

[0041] Figure 1 It is a flow chart for monitoring urban operation indicators based on multi-source heterogeneity.

[0042] Figure 2 It is a logical diagram for monitoring urban operation indicators based on multi-source heterogeneity. DETAILED DESCRIPTION

[0043] The embodiments of the present invention will be described below in conjunction with the accompanying drawings.

[0044] like Figure 1 and Figure 2 As shown, the present invention provides a method for monitoring urban operation indicators based on multi-source heterogeneity, comprising the following steps:

[0045] 1. Determine the city operation monitoring objectives and carry out monitoring indicator design;

[0046] 1-1. Determine monitoring indicators

[0047] Focusing on the economic governance scenarios of urban operations, we build indicators that can dynamically monitor economic operations. For example, in economic governance, we focus on data such as economic sustainable development, GDP growth rate, per capita disposable income, and growth rate of the number of registered enterprises.

[0048] Establish a monitoring indicator collection mechanism, establish data sharing channels with the Development and Reform Commission, Industry and Information Technology, Human Resources and Social Security, and Taxation Bureau, etc., regularly analyze these economic indicators, and evaluate the speed and stability of urban economic growth;

[0049] 1-2. Multidimensional analysis and construction of indicator library

[0050] According to different application scenarios, multi-dimensional analysis is used to design different types of indicators, including quantitative indicators, qualitative indicators, trend indicators and comparative indicators, to form a basic indicator library for daily monitoring of urban operations;

[0051] For indicators that can be analyzed quantitatively, determine specific numerical indicators, such as GDP, per capita disposable income, number of enterprises above designated size, etc., and use statistical methods and data analysis tools to conduct quantitative analysis of quantitative indicators;

[0052] For indicators that can be analyzed qualitatively, some descriptive indicators are formulated, such as the situation of economic development, the business environment for economic development, etc., and quantified by setting weights through the collection of multi-source data such as enterprise affairs handling, enterprise revenue, and enterprise headcount;

[0053] For trend analysis indicators, select some indicators that can reflect the trend of change, such as the trend of price index, the increase and decrease trend of employment, etc., and use time series analysis and other methods to monitor and analyze trend indicators;

[0054] For comparative analysis indicators, determine the indicators for comparison with other cities or regions, such as city competitiveness rankings, historical economic development levels, etc., collect relevant comparative data, and conduct horizontal comparison and analysis;

[0055] 1-3. Indicator coding and attribute definition

[0056] A unique indicator is determined for each record in the indicator library. The indicator library consists of three parts: indicator code, indicator name and indicator attribute;

[0057] Use a combination of numbers, letters or other characters to develop a unified coding rule to ensure that each indicator has a unique code for easy identification and management;

[0058] Define the name of the indicator, determine a clear and accurate Chinese name or English abbreviation of the indicator so that it can uniquely express a certain indicator and avoid using vague or ambiguous words;

[0059] Clarify the indicator attributes and define the connotation and extension of each indicator in detail, including indicator explanation, calculation method, logical relationship, accuracy, etc., to facilitate the maintenance and management of the indicator;

[0060] 1-4. Establish the indicator system management process

[0061] On the basis of planning and classifying indicators and data, a complete standardized management process for the indicator system is formulated, which can uniformly manage the changes of indicators to be released, review and release, indicator data access methods, online training, etc.;

[0062] Formulate a process for indicator changes, including submitting change applications, reviewing change content, and approving changes.

[0063] Establish a review and release mechanism to ensure the accuracy and reliability of indicators;

[0064] Clarify the data source and collection method for each indicator, such as collecting data through statistical reports, sensor monitoring, questionnaires, API interfaces, etc., and be able to configure scheduled tasks to automatically obtain or push data;

[0065] Carry out online training to improve data processing and analysis capabilities, and use machine learning, data mining and other technologies to conduct in-depth analysis and mining of data;

[0066] 1-5. Design a four-level indicator system

[0067] A four-level system is used to design indicators to ensure in-depth and detailed monitoring. The four-level system is designed: first-level fields (such as economic governance), second-level topics (such as industrial structure), third-level indicators (such as GDP growth rate), and fourth-level data items (such as GDP of the primary industry);

[0068] For each indicator, set the update frequency, responsible department, alarm threshold, target threshold and other key parameters of the data source corresponding to each indicator to ensure the timeliness and effectiveness of monitoring;

[0069] 2. Analyze monitoring indicators to determine data sources and conduct multi-source data aggregation management

[0070] 2-1. Data collection and verification

[0071] Determine the appropriate data collection method based on the data source corresponding to the monitoring indicator. For data sources with API interfaces, obtain data by calling the API; for data sources that can directly connect to the database, establish a database connection to extract data; for data sources that provide data in the form of files, set up a file transfer mechanism to receive data;

[0072] During the data collection process, write data verification rules and procedures to perform real-time verification on the collected data, such as checking whether the field is empty, whether the data type matches the expectation, whether the data value is within a reasonable range, etc. For data that does not meet the verification rules, record and mark it, and take corresponding processing measures, such as re-collection, data correction or notify the data source provider to make rectifications;

[0073] According to the needs of monitoring indicators and the update frequency of data sources, the frequency of data collection should be reasonably set. For indicators with high real-time requirements, such as traffic flow and air quality, real-time collection should be adopted; for some regularly updated data sources, such as statistical reports, the corresponding collection frequency should be set according to their release cycle, such as daily collection or weekly collection;

[0074] 2-2. Indicator items and data source association

[0075] Use statistical software (such as SPSS, SAS) and data mining tools (such as Weka, RapidMiner) to conduct correlation analysis and cluster analysis on monitoring indicators to identify repeatable and related indicators. Through correlation analysis, find out the linear or nonlinear relationship between indicators and data sources; through cluster analysis, group indicators and data according to similarity to find a set of indicators with strong correlation;

[0076] 2-3. Analysis and optimization of indicators

[0077] Conduct research and collect opinions, and conduct statistical analysis and text mining on the collected research data. Use statistical analysis methods to conduct descriptive statistics and correlation analysis on quantitative data; use text mining technology to extract keywords and perform semantic analysis on qualitative data. Optimize the selected indicators to form a set of practical and operational indicator systems;

[0078] 2-4. Tracing of indicator data

[0079] Record the entire process of data collection, processing and analysis, including data source, collection time, collection method, processing steps, analysis method, results, etc., to ensure that it can be integrated with the data collection, processing and analysis system to achieve automatic data recording and tracking;

[0080] Relying on tools such as data query, report generation and data visualization, various reports can be automatically generated according to user needs, such as data summary reports, trend analysis reports, etc., to facilitate managers to query, analyze and verify data;

[0081] 2-5. Update of indicator data

[0082] Clarify the update time interval, update content and scope, and formulate a data source update strategy that combines regular updates and on-demand updates based on the characteristics of the data source and the needs of monitoring indicators;

[0083] For data sources with fixed update cycles, such as GDP and per capita disposable income, a regular update method is adopted; for data sources with frequent changes, such as sensor data and social media data, an on-demand update method is adopted;

[0084] By setting up scheduled tasks or triggers, data update operations are automatically triggered when the data source changes;

[0085] 2-6. Automatic labeling and classification

[0086] Automatically label and classify multi-source data using machine learning algorithms such as classification and clustering algorithms and natural language processing technologies such as lexical analysis, syntactic analysis, and semantic understanding;

[0087] Establish a list of associated mappings between indicators and data resources, and clarify the data source and data field corresponding to each indicator;

[0088] Through the visual configuration management of indicator items, users can set the label attributes of indicator items, such as indicator name, annotation method, display format, etc.

[0089] 2-7. Data aggregation and governance

[0090] Use ETL tools and data warehouse technology to manage the task of aggregating data. Through ETL tools, extract, clean, transform and load heterogeneous multi-source data, and integrate data from various data sources into the city operation management database;

[0091] During the data aggregation process, implement data quality monitoring, data flow monitoring and system performance monitoring. By setting monitoring indicators and thresholds, timely discover data quality problems, data flow anomalies and system performance bottlenecks;

[0092] Establish an exception handling mechanism to promptly respond to and handle abnormal situations that occur during data collection, processing, and transmission. For example, when data quality does not meet requirements, clean and correct the data; when data traffic is too large or system performance decreases, take optimization measures or allocate resources;

[0093] 2-8 Data Trust Space

[0094] Establish a role- and permission-based access control mechanism to manage users and data by category, and assign corresponding data access permissions to users of different roles;

[0095] Use encryption algorithms such as AES and RSA to encrypt the aggregated data to protect the confidentiality and integrity of the data;

[0096] Implement sensitive data automatic identification technology to automatically identify and mark sensitive information in the data, and take corresponding encryption and desensitization measures, such as data replacement and data obfuscation;

[0097] Use data security watermark technology to mark and track data to prevent illegal copying and dissemination of data;

[0098] 3. Build a city operation indicator monitoring model with adaptive capabilities

[0099] 3-1. Data cleaning and analysis

[0100] De-duplicate the collected data to remove duplicate data records and ensure the uniqueness of the data;

[0101] By setting reasonable data ranges and logical rules, erroneous data can be identified and corrected, and erroneous values ​​in the data, such as outliers and illogical data, can be cleaned;

[0102] For unstructured data, such as text data, natural language processing (NLP) technology is used for parsing, including lexical analysis, syntactic analysis, semantic understanding, etc., to convert text data into structured or semi-structured form;

[0103] For unstructured data such as images, use image recognition technology (OCR) to convert them into text information, or extract feature information of the image and convert it into structured data that can be used for analysis;

[0104] 3-2. Data Dimensionality Reduction and Standardization

[0105] Select appropriate dimensionality reduction techniques to map high-dimensional data to low-dimensional space, reduce the dimension of the data, and reduce computational complexity. Principal component analysis (PCA) or t-distributed stochastic neighbor embedding (t-SNE) can be used to process the cleaned and parsed data and remove redundant information.

[0106] The Z-Score standardization method is used to standardize the data, calculate the mean and standard deviation of each data point, and convert the data value into a standard score to make the data of different dimensions comparable and converted to the same dimension;

[0107] 3-3. Define monitoring indicator prototype

[0108] Formulate the calculation formula for the indicator and clarify the data fields and calculation methods involved in calculating the indicator value;

[0109] Determine the data objects related to each indicator, including data sources (such as databases, file systems, etc.), data fields (such as specific table fields), and data types (such as numeric, character, date, etc.);

[0110] 3-4. Preliminary construction of monitoring model

[0111] Choose a suitable algorithm framework based on the characteristics of the indicator and the distribution of the data. If the data has a linear relationship, you can choose a linear regression algorithm; if the data has a complex nonlinear relationship, you can choose a neural network algorithm; if classification and decision-making are required, you can choose a decision tree algorithm, etc.

[0112] Use the selected algorithm framework and tools to build the original model of monitoring indicators based on the indicator library prototypes and data objects;

[0113] Set the initial parameters of the model. The initial parameters can be set based on empirical values ​​or random initialization methods, such as the initial values ​​of coefficients in linear regression, the splitting criteria and depth limit in decision trees, the number of layers and neurons in neural networks, etc.

[0114] 3-5. Model training and parameter optimization

[0115] According to business needs and data characteristics, set reasonable parameter weights and indicator thresholds for the model. The weights are used to adjust the impact of different features on the model output, and the thresholds are used to determine the boundaries of classification or decision-making.

[0116] Divide the data into training set, validation set and test set, use historical data or simulated data to train the model, and optimize the performance of the model by adjusting the model parameters;

[0117] 3-6. Building a knowledge graph

[0118] Use machine learning and natural language analysis techniques to analyze and understand data and extract entity, relationship and attribute information from the data;

[0119] Use graph databases or knowledge graph building tools to represent complex associations and relationships between data, and build the extracted entities, relationships, and attribute information into a knowledge graph;

[0120] Combine the constructed knowledge graph with the monitoring indicator model, and optimize the structure and parameters of the model through the relationship information in the knowledge graph;

[0121] 3-7. Model parameter tuning

[0122] Select the Adam optimizer to tune the model parameters. During the iterative training process, the weights and bias items of the model are gradually adjusted according to the optimizer's algorithm and objective function;

[0123] Monitor parameter changes and model performance during the tuning process, and adjust optimizer parameters and training strategies in a timely manner based on training results to ensure that the model can quickly converge to the optimal solution;

[0124] 3-8. Relationship extraction and mapping

[0125] Use rule-based methods to extract entities and relationships related to indicators from text data according to predefined rules and patterns; use statistical-based methods to automatically identify relationship information related to indicators through statistical analysis and pattern discovery of data;

[0126] Map the extracted relationship information with the indicator model, establish a relationship mapping table or relationship map, clarify the correspondence between the indicator and the relationship information, and facilitate query and use;

[0127] 3-9. Automatic extraction of related data

[0128] Design the extraction process of associated data based on the relationship mapping table or relationship graph, determine the source of data, extraction methods and steps, data conversion and processing flow, etc.;

[0129] Use workflow engines or process management tools to automatically arrange and manage the extraction process, and implement process definition, configuration, and execution, including task scheduling, resource allocation, process monitoring, and control.

[0130] 3-10. Model publishing and visualization

[0131] Deploy the verified and optimized indicator model to the production environment to realize online operation and service provision of the model;

[0132] Establish a monitoring and maintenance mechanism for the model, evaluate and update the model regularly to ensure its performance and accuracy;

[0133] Use visualization technologies such as data dashboards and 3D simulation to display and output model calculation results;

[0134] 4. According to the above process, the design of operation monitoring indicators for key areas of urban operation such as industrial economy, urban governance, green ecology, happiness and livability shall be carried out in a hierarchical and classified manner, and the operating steps are similar.

[0135] In a specific implementation, the present application provides a computer storage medium and a corresponding data processing unit, wherein the computer storage medium can store a computer program, and when the computer program is executed by the data processing unit, the invention content of a method for monitoring city operation indicators based on multi-source heterogeneity provided by the present invention and some or all steps in each embodiment can be executed. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0136] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present invention can be implemented by means of computer programs and their corresponding general hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention are essentially or partly contributed to the prior art can be embodied in the form of a computer program, i.e., a software product, which can be stored in a storage medium and includes several instructions for enabling a device including a data processing unit (which can be a personal computer, a server, a single-chip microcomputer, a MUU or a network device, etc.) to execute the methods described in various embodiments of the present invention or certain parts of the embodiments.

[0137] The present invention provides a method for monitoring city operation indicators based on multi-source heterogeneity. There are many methods and ways to implement the technical solution. The above is only a preferred implementation of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the protection scope of the present invention. All components not specified in this embodiment can be implemented by existing technologies.

Claims

1. A method for monitoring urban operation indicators based on multi-source heterogeneity, characterized in that: The following steps are involved: Step 1: Determine the city operation monitoring objectives and design monitoring indicators; Step 2: Analyze monitoring indicators to determine data sources and conduct multi-source data aggregation and management; Step 3: Build a city operation indicator monitoring model with adaptive capabilities; Step 4: Carry out monitoring and application of urban operation indicators in a hierarchical and classified manner.

2. The method for monitoring urban operation indicators based on multi-source heterogeneity according to claim 1 is characterized in that: Step 1 includes: Step 1-1: Conduct a panoramic scan of urban operations around the economic governance, urban governance and public safety scenarios of urban operations, and construct indicators that can dynamically monitor urban operations; Step 1-2: Design different types of indicators using multidimensional analysis for different application scenarios to form a basic indicator library; Step 1-3, determine a unique indicator for each record in the basic indicator library. The management definition of the indicator consists of three parts: indicator code, indicator name and indicator attribute; Steps 1-4: Develop a complete standardization management process for the indicator system, which can uniformly manage indicator change requirements, review, release, and indicator data entry changes; Steps 1-5: The indicator design adopts a four-level system, with first-level fields, second-level topics, third-level indicators, and fourth-level data items. The indicator items at least include indicator name, indicator unit, update frequency, responsible department, alarm threshold and target threshold.

3. The method for monitoring urban operation indicators based on multi-source heterogeneity according to claim 2 is characterized in that: Step 2 includes: Step 2-1: When selecting indicators, consider the scientificity and convenience of data collection, and clarify the data source used for each indicator; determine the data collection method based on the data source corresponding to the monitoring indicator, and perform real-time verification of the collected data during the data collection process; set the frequency of data collection based on the needs of the monitoring indicator and the update frequency of the data source; Step 2-2, analyze the repeatability and correlation between indicators, and screen or aggregate multiple indicators with repeatability and correlation; Step 2-3: Conduct research and collect opinions, conduct statistical analysis and text mining on the collected research data, optimize the selected indicators, and form a complete and operational indicator system.

4. The method for monitoring urban operation indicators based on multi-source heterogeneity according to claim 3 is characterized in that: Step 2 also includes: Step 2-4: Establish a data traceability mechanism to obtain the full chain information from the source to the mapping indicator, drill down to the basic data layer by layer according to the indicator level, and obtain the situation of each indicator under the jurisdiction of the superior indicator, ensuring that the indicator is traceable and the data can be checked; Steps 2-5: Establish a data source update mechanism, clarify the update cycle, responsible department, and update time; at the same time, establish a system automatic inspection mechanism to ensure the real-time freshness of indicator data.

5. The method for monitoring urban operation indicators based on multi-source heterogeneity according to claim 4 is characterized in that: Step 2 also includes: Step 2-6: According to the design indicator classification, a list of urban operation indicator data resources is constructed, which can automatically adapt the labels of the collected multi-source data to ensure accurate data acquisition for different indicator items; Steps 2-7: Manage the tasks of aggregating data, form tasks and automatically perform data aggregation, realize the extraction, cleaning, loading, conversion, and transmission of heterogeneous multi-source data, and complete the management and monitoring of the collected and aggregated data; Steps 2-8: Set security rules for the aggregated data, automatically identify sensitive data, encrypt and desensitize data, classify and grade data, control access permissions, authorize and approve data, and implement data security watermarking to ensure data security.

6. The method for monitoring urban operation indicators based on multi-source heterogeneity according to claim 5 is characterized in that: Step 3 includes: Step 3-1: According to the designed indicator items and the aggregated multi-source data items, automatically construct the structured data and unstructured data, conduct multi-dimensional data analysis, and perform dimensionality reduction and standardization on the multi-source heterogeneous data; Step 3-2: Establish a sample library, preliminarily establish the original model of monitoring indicators based on the prototype definition of the indicator library and the data objects involved, set the model weights, thresholds and evaluation criteria, and carry out model training; Step 3-3, perform parameter tuning and rule adaptation on the constructed indicator model to provide decision support for the accuracy of corresponding indicator monitoring; Step 3-4: Carry out automated orchestration and management of the extraction process of associated data corresponding to different indicators, including relationship extraction, relationship mapping, relationship filtering and automatic extraction of associated data; Steps 3-5, publish and apply the indicator model online, combine thematic business data and two- and three-dimensional geographic data, organically integrate key data indicators with the city model, and form a panoramic view of the city operation.

7. The method for monitoring urban operation indicators based on multi-source heterogeneity according to claim 6 is characterized in that: Step 3-2 includes: Define the monitoring indicator prototype according to business needs and data characteristics, including the indicator name, description and calculation formula; determine the data objects related to the indicator, including data sources, data fields and data types; Based on the monitoring indicator prototype and data objects, select linear regression, decision tree, and neural network algorithm frameworks, preliminarily build the original model of the monitoring indicator, and set the initial parameters of the model; According to business needs and data characteristics, set weights, thresholds, and evaluation criteria for the model, and use historical data or simulated data to train the model to adjust the model parameters.

8. The method for monitoring urban operation indicators based on multi-source heterogeneity according to claim 7 is characterized in that: Step 3-3 includes: Use machine learning and natural language analysis techniques to analyze and understand data and extract entity, relationship and attribute information from the data; Use graph databases or knowledge graph building tools to represent complex associations and relationships between data, and build the extracted entities, relationships, and attribute information into a knowledge graph; Combine the constructed knowledge graph with the monitoring indicator model, and optimize the structure and parameters of the model through the relationship information in the knowledge graph; Gradient descent, stochastic gradient descent, and Adam optimizer machine learning algorithms are used to tune the parameters of the monitoring indicator model. During the iterative training process, the weights and bias items of the model are gradually optimized.

9. The method for monitoring city operation indicators based on multi-source heterogeneity according to claim 8 is characterized in that: Steps 3-4 include: Use a variety of relationship extraction methods to automatically extract relationship information related to corresponding indicators from multi-source heterogeneous data, and map the extracted relationship information with the urban indicator model to establish a relationship mapping table or relationship map; According to the relationship mapping table or relationship graph, the associated data extraction process is automatically orchestrated and managed, including process design, task scheduling and error handling, and the associated data related to the corresponding indicators are automatically extracted.

10. The method for monitoring city operation indicators based on multi-source heterogeneity according to claim 9 is characterized in that: Step 4 includes: For the areas of industrial economy, urban governance, green ecology and happy and livable city operations, carry out steps 1 to 3 in a hierarchical and classified manner to carry out the design of operation monitoring indicators.

Images